When trying and failing to install a 4.4 IPv6 or dual-stack cluster on Azure (following the directions from https://github.com/openshift/installer/pull/2847), the installer then fails to run must-gather because it uses the wrong IP address.
Verified this bug using installer from payload 4.4.0-0.nightly-2020-02-15-024625 $ ./openshift-install version ./openshift-install v4.4.0 built from commit a627412646a322daa6596bdb5bcefe8c64690439 release image registry.svc.ci.openshift.org/ocp/release@sha256:5de81fdc8361b00a6b46dfa822f52e7495f8555cbc88d99b4b634869c1f9ddb1 After a 4.4 ipv6 single-stack installation on Azure failed(which is expected for now), use the installer to gather debugging logs. $ ./openshift-install gather bootstrap --key ./x.pem --bootstrap 13.89.107.24 --master 10.0.0.5 INFO Pulling debug logs from the bootstrap machine INFO Bootstrap gather logs captured here "log-bundle-20200215055500.tar.gz" $ ./openshift-install gather bootstrap --key ./x.pem --bootstrap 2603:1030:b:3::48 --master 10.0.0.5 INFO Pulling debug logs from the bootstrap machine INFO Bootstrap gather logs captured here "log-bundle-20200215055739.tar.gz" $ ./openshift-install gather bootstrap --key ./x.pem --bootstrap 2603:1030:b:3::48 --master fd00::5 INFO Pulling debug logs from the bootstrap machine INFO Bootstrap gather logs captured here "log-bundle-20200215060021.tar.gz" $ ./openshift-install gather bootstrap --key ./x.pem --bootstrap 13.89.107.24 --master fd00::5 INFO Pulling debug logs from the bootstrap machine INFO Bootstrap gather logs captured here "log-bundle-20200215072937.tar.gz"
Also found there's some problem when using master's ipv6 address for gathering logs, filed https://bugzilla.redhat.com/show_bug.cgi?id=1803321 to track.
(In reply to Gaoyun Pei from comment #2) > After a 4.4 ipv6 single-stack installation on Azure failed(which is expected > for now), use the installer to gather debugging logs. > > $ ./openshift-install gather bootstrap --key ./x.pem --bootstrap > 13.89.107.24 --master 10.0.0.5 Sorry, the problem was that it was detecting the wrong IP address when it tried to *automatically* run must-gather after a failed install. If you are manually running must-gather and telling it explicitly what IP to use then that's not testing the fix.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:0581